PROJECT SUMMARY Despite the availability of continuous glucose monitors (CGM) and insulin pumps, relatively few persons with type 1 diabetes (T1D) meet recommended glycemic targets; furthermore, rates of severe hypoglycemia remain unacceptably high. These suboptimal outcomes are especially true for children, adolescents, and young adults with T1D. Automated ?closed-loop? (CL) insulin delivery systems (so-called ?artificial pancreas? [AP]) promise to improve glucose control and reduce hypoglycemia while preserving quality of life. Early generations of CL systems have shown beneficial effects on glycemic outcomes, but their design and form factors may not be optimal for user satisfaction and continued use: most systems use a tubed insulin pump and smartphone controller device, two critical system components that challenge long-term usability. Ultimate universal adoption of these technologies will depend not only on their effectiveness and safety, but also on ?human factors? related to ease and comfort of use, as perceived by the patient and family. Further, most control algorithms do not allow for customization of glycemic parameters or alerts and alarms. We believe that a completely ?on-body? system has distinct advantages, including: function during activities such as exercise and showering; greater user satisfaction with fewer disparate devices; and additional benefits can be realized with customizable, remote alert messaging to smartphones of users and/or family members. These improvements will likely maximize glycemic and quality of life benefits. This application presents an iterative series of studies that will enable the development of a completely on-body CL system based on the validated Model Predictive Control (MPC) AP strategy developed by Dassau and Doyle that is enhanced with an adaptive learning module and customizable, user-programmable alert messages based on individual needs. Our 3 specific aims [targeting effective and safe control of BG levels, user satisfaction, and quality of life] follow: Specific Aim 1: Develop and evaluate an automated insulin delivery closed-loop system with a fully ?on-body? ecosystem for use in pediatric, adolescent, and young adult subjects with T1D. An on- body CL system will be developed and evaluated in a series of feasibility studies, starting in young adults and moving to adolescents and then younger children. Specific Aim 2: Design a customizable, remote alert messaging system for use with the on-body CL system. We will assess patient and family preferences regarding usability, acceptability, and alert/alarm settings through focus groups/semi-structured interviews and surveys of study subjects and family members, with the goal to develop a platform that enables alert customization Specific Aim 3: Develop an adaptive learning module for use with the CL system. An adaptive learning control algorithm designed to improve performance over time will be evaluated in a series of home- based effectiveness studies, starting in young adults and moving to adolescents and then younger children.